Pubblicazione scientifica

Fast Visual Tracking Based on Convolutional Networks

FILONE TECNOLOGICO	TEMA
produzione e raccolta

RIASSUNTO

Recently, an upsurge of deep learning has provided a new direction for the field of computer vision and visual tracking. However, expensive offline training time and the large number of images required by deep learning have greatly hindered progress. This paper aims to further improve the computational performance of CNT which is reported to deliver 5 fps performance in visual tracking, we propose a method called Fast-CNT which differs from CNT in three aspects: firstly, an adaptive k value (rather than a constant 100) is determined for an input video; secondly, background filters used in CNT are omitted in this work to save computation time without affecting performance; thirdly, SURF feature points are used in conjunction with the particle filter to address the drift problem in CNT. Extensive experimental results on land and undersea video sequences show that Fast-CNT outperforms CNT by 2~10 times in terms of computational efficiency.

DATA

Data di pubblicazione: 24/07/2018

AUTORI

HUANG REN-JIE
TSAO CHUN-YU
KUO YI-PIN
LAI YI-CHUNG
LIU CHI CHUNG
TU ZHE-WEI
WANG JUNG-HUA
CHANG CHUNG-CHENG

ENTE DI AFFERENZA

DEPT ELEC ENG NATL TAIWAN OCEAN UNIV KEELUNG CITY 20224 TAIWAN 20553001 MAIL NTOU EDU TW R JH 10453016 MAIL NTOU EDU TW C YT K8130488 GMAIL COM Y CL FREE TOUCH TIM HOTMAIL COM CCL JJAY6TWTW GMAIL COM Z WT CCCHANG MAIL NTOU EDU TW C CC
SHIP & OCEAN IND R & D CTR SOIC NEW TAIPEI CITY 25170 TAIWAN S880062 GMAIL COM

RIVISTA

Sensors (Basel, Switzerland), 18(8), 2405

IMPACT FACTOR

3,510

indietro